Tāmaota operates on a radically new basis of crowd sourcing: Contributions to the company are directly converted in equity whenever they proved valuable. So, what does it mean?
Tāmaota rewards data contribution by means of the expected utility of the new data. For a deep dive into the theory behind, start with Wikipedia and follow respective references.
In a nutshell, consider the following situation: Alice and Bob are providing datasets A and B. Both datasets have same number of data points and are mutually exclusive. A model can be created from simply summing up A+B and is sold to a customer. What would you consider a fair split of the value Alice and Bob contributed respectively? We would consider a 50:50 split fair, and so does the expected utility theory.
Now consider Charlie is adding dataset C, but it happens that this dataset is actually identical to existing dataset A. Here, Charlie is not adding any new information to Tāmaota’s data base (besides the fact that he is somehow validating dataset A by resubmission). The expected utility of C is zero, and such is Charlie’s equity gain in case of model A+B being sold.
These toy examples illustrate the concept in brief. We are about preparing a detailed technical paper on the applied rule set and will publish it here in short time.